The invention pertains to the field of fault tolerant arrays of hard disks that are known as Redundant Arrays of Inexpensive Disks (RAID). More specifically the invention relates to the management and correction of undetected read errors in a RAID system that are discovered when a disk failure requires the replacement and build-up of a disk other than that containing the undetected read error.